CN105869085A - Transcript inputting system and method for processing images - Google Patents
Transcript inputting system and method for processing images Download PDFInfo
- Publication number
- CN105869085A CN105869085A CN201610186428.9A CN201610186428A CN105869085A CN 105869085 A CN105869085 A CN 105869085A CN 201610186428 A CN201610186428 A CN 201610186428A CN 105869085 A CN105869085 A CN 105869085A
- Authority
- CN
- China
- Prior art keywords
- image
- point
- data input
- paper
- student
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/20—Education
- G06Q50/205—Education administration or guidance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
- G06V10/225—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition based on a marking or identifier characterising the area
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/14—Image acquisition
- G06V30/146—Aligning or centring of the image pick-up or image-field
- G06V30/1475—Inclination or skew detection or correction of characters or of image to be recognised
- G06V30/1478—Inclination or skew detection or correction of characters or of image to be recognised of characters or characters lines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/24—Character recognition characterised by the processing or recognition method
- G06V30/248—Character recognition characterised by the processing or recognition method involving plural approaches, e.g. verification by template match; Resolving confusion among similar patterns, e.g. "O" versus "Q"
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Educational Technology (AREA)
- Educational Administration (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- Health & Medical Sciences (AREA)
- Economics (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a transcript inputting system and a method for processing images. The transcript inputting system includes an acquiring unit, a transcript inputting terminal, a firewall, a router, a center server and a client. The method includes training a handwritten form digital identification classifier, acquiring original color images, conducting processing, identifying and analyzing on images, and displaying processing results. The beneficial effects of the method are: transcript inputting operation becomes simplified, efficient and intelligentized, and exam result analysis that is detailed, accurate and scientific can be provided for teachers and school leadership, and the system and the method are applicable to any exam in the field of education.
Description
Technical field
The invention belongs to digital education technical field, a kind of method relating to Students'Transcript Input and image procossing.
Background technology
Along with computer technology, big data analysis and the high speed development of artificial intelligence technology, education and instruction works the most no longer
It is traditions of the past modes, but pursues highly efficient, science, the education and teaching idea in forward position.Current in order to guarantee fairness,
Objectivity, either to take an advanced study study further, also be intended to examine civil servant or public institution etc., is required for by examination people
Member selects.Taking an exam the most frequently in exercise and interim examination detection for a long time is all to exist.And examine
Examination data, as the important evidence of embodiment Students ' Learning situation, can guide the learning direction that student is follow-up, the most also can be for religion
Shi Gaijin teaching provides help.Therefore collection and statistical analysis for examination data are significant.
At present, a lot of research institutions and enterprise, in order to collect examination data easily, propose a kind of online testing
Form.Such as Chinese patent discloses one " examination system " (publication number: CN 103617584 A), and this system includes system
Manage module, examination management module, individual's test modules, blank pipe is managed module and statistical analysis module.Although above-mentioned patent can
To get examination data the most easily, and do Knowledge Relation, give detailed statistical data analysis.But this germline
Unite in current K12 educates is to be difficult to promote.Limit in view of economic condition, the problem holding online testing intricate operation,
Whole school's unified examination realizes the problems such as the restriction of difficulty, and the examination every time of middle and primary schools is desirable that it is impossible online.Cause
This sees with regard to present circumstances, and interior, school still can take an exam with papery paper form.
Summary of the invention
The technical problem to be solved is to provide a kind of at the base not changing current Faculty and Students examination custom
On plinth, achievement is made to step on system simple operation, efficient and intelligentized Students'Transcript Input and the method for image procossing.
Be the technical scheme is that a kind of Students'Transcript Input by realizing technique scheme, it include collecting unit,
Data Input client, fire wall, router, central server and user terminal；
The input of described collecting unit obtains paper image；The output of described collecting unit terminates described Data Input client
Respective input；Described Data Input client is connected with the Internet；
Described central server is connected with the Internet through router, fire wall；
Described user terminal is connected with the Internet.
Described collecting unit is scanner, high photographing instrument, photographing unit, video camera or mobile phone camera.
Described Data Input client is computer, panel computer or mobile phone.
Described central server is computer；Described user terminal is computer.
Utilizing the method that described Students'Transcript Input carries out image procossing, it comprises the steps:
(1) training of Handwritten Digital Recognition grader is carried out: by collection and the screening of handwriting digital sample, it is thus achieved that effectively
Handwriting digital sample, the method utilizing support vector machine, train and generate identify handwriting digital svm grader；
(2) utilize described collecting unit collection with the paper original color image of scoring papery of volume head；
(3) Data Input client obtains the paper original color image of scoring papery with volume head；Data Input client
By gray processing processing method, the described paper original color image of scoring papery with volume head is converted into gray level image；
(4) described gray level image is converted into bianry image by Data Input client；
(5) Data Input client utilizes Gassian low-pass filter method that described bianry image carries out denoising for the first time, obtains once
Denoising bianry image；
(6) a described denoising bianry image is found datum line by hough conversion by Data Input client, and according to benchmark
Line computation angle of inclination, finally carries out slant correction to a denoising bianry image, obtains correcting a rear denoising bianry image；
(7) Data Input client carries out connected component labeling to correcting a rear denoising bianry image, finds all of connected region
Territory, and according to the information of connected region size, connected region is screened, distracter the least for area is got rid of, obtains
Connected region after screening；
(8) Data Input client utilizes positional information that connected region after described screening carries out the coarse positioning of paper volume head,
Head image is rolled up to coarse positioning examination papers；
(9) Data Input client carries out horizontal direction, vertical direction projection respectively to coarse positioning examination papers volume head image, checks
Whether meet volume head table features；The part meeting volume head table features is defined as rolling up the exact position of head；
(10) Data Input client carries out cutting according to the exact position of volume head to correcting a rear denoising bianry image, extracts
Go out accurately volume head image；
(11) Data Input client carries out horizontal direction, vertical direction projection respectively to accurately volume head image, calculates form
Coordinate position；
(12) Data Input client carries out vertical segmentation to image, it is thus achieved that little topic score graph table images；
(13) Data Input client carries out the segmentation of horizontal direction to little topic score graph table images, it is thus achieved that little topic mark each
The image of bit digital；
(14) according to the paper structural information obtained from central server, segmentation result is verified by Data Input client,
Authentication image segmentation is the most accurate；
(15) Data Input client utilizes Gaussian low pass wave method that each handwriting digital image is carried out denoising, is gone
Handwritten form digital picture after making an uproar；
(16) Data Input client carries out image enhaucament by morphologic expanding method to handwritten form digital picture after making an uproar, and obtains
Obtain enhancement mode handwriting digital image；
(17) by being inwardly indented the method with morphological reconstruction, enhancement mode handwriting digital image is entered by Data Input client
Row is removed frame and is processed；
(18) the enhancement mode handwriting digital image removing frame is carried out numeral knowledge by svm grader by Data Input client
, do not finally give the fractional data information of per pass exercise question, and described fractional data information be transferred to described central server；
(19) described central server carries out data process to described fractional data information, obtains analysis processing result；
(20) described user terminal and is divided from the fractional data information of described central server acquisition per pass exercise question by the Internet
Analysis result, and fractional data information and the analysis processing result of described per pass exercise question is shown by user terminal displays device.
Above-mentioned image processing method, utilizes Highcharts by described analysis processing result with form, rectangular histogram, curve chart
Or the form that word describes shows user.
The data of described fractional data information process and include that average mark statistics, item difficulty prediction curve and item difficulty are real
Survey curve and examination diagnosis report；
Described average mark statistics is in units of class, by suing for peace, every student's score in described class then divided by always
Number, obtains the average mark of this class, and expression formula is:
；Wherein, T is this class of answer number,Achievement for i-th student.
It is as follows that described item difficulty prediction curve and item difficulty measured curve generate method:
The formula calculating item difficulty value is:, wherein P is examination question Awareness percent,, it is right that R is that this examination question does
Number, T for participate in answer class size；
With topic number as abscissa, item difficulty value is vertical coordinate, marks corresponding coordinate points in coordinate, and according to topic priority
It is sequentially connected with coordinate points with smoothed curve, i.e. obtains item difficulty prediction curve.
After Data Input completes, the true Awareness percent of per pass examination question is RP, calculates per pass examination question true difficulty value RV, table
Reaching formula is；With topic number as abscissa, the true difficulty value of examination question is vertical coordinate, marks corresponding in coordinate
Coordinate points, and it is sequentially connected with coordinate points according to topic priority smoothed curve, i.e. obtain item difficulty measured curve.
Described examination diagnosis report includes the diagnosis to paper, rate of losing points diagnosis, the examining of every student knowledge Grasping level
Disconnected, student ability point diagnosis；
A, diagnosis to paper include for distinguishing the examination question discrimination of horizontal student at all levels, Degree of difficulty of test paper, knowledge point
Collocation stresses with ability point；
Examination question discrimination refers to the examination question differentiation degree to examinee's practical level；The computing formula of examination question discrimination is
D=PH-PL
Wherein D is examination question discrimination, and PH, PL are respectively the packet of examination question height and the Awareness percent of low packet examinee, and high packet is just answered
Rate formula is:, wherein RH height packet do to number, TH is high packet number；Low packet Awareness percent formula is:, wherein the low packet of RL do to number, TL is low packet number；
Described Degree of difficulty of test paper is weighed by examination question average degree of difficulty；Described average degree of difficulty is all item difficulty values
Actual measurement meansigma methods；
The formula of average degree of difficulty A is:；WhereinBeing the i-th problem purpose difficulty value, q is examination question in paper
Sum；
Knowledge point collocation is the distribution proportion listing each knowledge point, checks that the distribution of knowledge point is the most averagely or side
Weight；Distribution proportion is to obtain divided by paper total score according to the mark shared by each knowledge, and distribution proportion formula is:；WhereinFor the mark shared by kth knowledge point；N is the number of knowledge point,For i-th knowledge
The shared mark of point；
Ability point collocation is the distribution proportion listing each ability point, checks that the distribution of ability point is the most averagely or side
Weight；Ability point distribution proportion computing formula is similar with knowledge point ratio distribution formula, i.e. removes with the mark shared by each ability point
With paper gross score；Distribution proportion formula is:；WhereinFor the mark shared by kth ability point；m
For the number of ability point,Mark shared by i-th ability point；
The purpose of b, rate of losing points diagnosis is to count whole class to lose points the higher exercise question of rate, and advises that teacher's emphasis explains this
The content of class exercise question；The rate of wherein the losing points problem definition more than 40% is the higher examination question of rate of losing points, and the rate formula of losing points is, wherein L is the number that this examination question there is no full marks, and T is this class of answer number；
C, the diagnosis of student knowledge Grasping level
Per pass exercise question has corresponding knowledge point, is it is recognized that student is to knowledge according to student's per pass exercise question scoring event
The Grasping level of point；The diagnosis of student knowledge Grasping level is the scoring rate counting each knowledge point；Wherein knowledge point score
Rate formula is:, whereinThe mark obtained by this knowledge point of i-th classmate of this class, KF knows for this
Know the gross score that point is shared in paper；T is this class of answer number；
D, student ability point diagnoses
Per pass exercise question has corresponding ability point, is it is recognized that student is to ability according to student's per pass exercise question scoring event
The Grasping level of point；The diagnosis of each ability Grasping level of student is the scoring rate counting each ability point；Computing capability point
Scoring rate formula is:；The mark obtained by this ability point of i-th classmate of this class, AF is this energy
The gross score that force is shared in paper；T is this class of answer number.
The invention has the beneficial effects as follows: be to make achievement step on the simplification of system operation change, high efficiency, intellectuality, and can
There is provided in detail for teacher and school leaders, accurately, scientifically examination result analysis；Any be applicable to education sector of the present invention
Examination.
Accompanying drawing explanation
Fig. 1 is Students'Transcript Input theory diagram of the present invention.
Fig. 2 is Data Input client operational effect figure sectional drawing.
Fig. 3 is image processing algorithm flow chart.
Fig. 4 is for stepping on system paper achievement list figure.
Fig. 5 is average mark block diagram.
Fig. 6 is item difficulty prediction and measured curve figure.
Fig. 7 is paper diagnostic graph.
Fig. 8 is rate diagnostic graph of losing points.
Fig. 9 is student knowledge Grasping level diagnostic graph.
Figure 10 is student ability Grasping level diagnostic graph.
Detailed description of the invention
Below in conjunction with Fig. 1-10 and embodiment, the present invention is illustrated.
In order to each examination data of simple, efficient, intelligent collection, the present invention proposes a kind of based on digital picture
The Students'Transcript Input processed.Papery paper information is entered in system by this system automatically, and enters examination data information
Go detailed statistics and analysis, it is possible to rapidly, easily, scientifically help teachers ' analysis examination result.
For achieving the above object, the technical solution adopted in the present invention is as follows:
Image capture device and Data Input client are attached (such as, Data Input client and image acquisition in mobile phone
Except equipment is integrated), enable Data Input client-side program to get the information of image capture device, and recorded by achievement
Enter client image capture device is controlled, papery paper collection is become paper image.
In Data Input client, first image processing module carries out image procossing to paper image, by image two-value
Changing, subregion coarse positioning filled out by Morphological scale-space, connected domain analysis, paper, paper is filled out subregion and is accurately positioned, image cut, fills out
The steps such as subregion image slant correction extract paper and fill out subregion tabular drawing picture, by throwing filling out a point tabular drawing picture
Shadow, cuts out the score chart picture of per pass exercise question according to projection information.Then the picture recognition module identification of Data Input client
Go out the mark of corresponding exercise question, and on client end interface, show recognition result.Final result typing client passes through Web
Service technology will identify that correct mark uploads to central server.
Result statisticses and analysis end obtains data from central server, system pass through counting statistics, by analysis result with form,
The forms such as rectangular histogram, curve chart and word description show user.Main contents include: the statistics of average mark, and item difficulty is pre-
Survey and measured curve, diagnosis report of taking an examination.Wherein diagnosis report includes that the diagnosis to paper, rate of losing points diagnosis, every student know
Know the diagnosis of Grasping level, the diagnosis of student ability point.
In order to more clearly understand the technological means of the present invention, and can be practiced according to the content of description, with
Under with a preferred embodiment and accompanying drawing, technical scheme is further described.
Students'Transcript Input refers to Fig. 1, and image capture device described in the present embodiment is high photographing instrument.Data Input client end
It is deployed on the desktop computer with windows system.High photographing instrument is connected with desktop computer by data wire.The most chartered user runs
Data Input client-side program, carries out login authentication by user name, password, identifying code.Data Input client operational effect
Refering to Fig. 2.
Data Input client includes paper preview pane, rolls up head form preview pane, mark identification display box and current the most
Step on system paper achievement list display window.During teacher steps on point, first paper is placed in the scope captured by high photographing instrument, logical
Cross paper preview pane and guarantee that paper is placed accurately, ensure that volume head form is at the center of paper preview pane as far as possible.Then by tapping
" space " key or left mouse button in keyboard are clicked on " taking pictures " button control high photographing instrument and are carried out paper image acquisition.Obtain image
Volume head is stepped on point form and is split from original paper image by rear client application image processing techniques, and result is shown
Show in volume head form preview pane.Mark recognition result is shown point box after recognition by final result typing client automatically
In.Image processing algorithm flow chart refers to Fig. 3, and main contents are as follows:
Before paper image procossing and identification, first have to carry out the training of Handwritten Digit Classification device.By the collection of numeral sample,
The screening of sample, it is thus achieved that effective handwriting digital sample, by supporting the method to vector machine (svm), trains and generates number
The svm grader of word identification.
The first step: gather paper original color image by image capture devices such as high photographing instrument.
Second step: processed by gray processing and original color image is converted into gray level image.
3rd step: apply local auto-adaptive binarization method or first passing through sobel operator edge extracting reapplies ostu
Binarization method, is converted into bianry image by gray scale paper image.
4th step: application Gassian low-pass filter carries out denoising for the first time to the bianry image of paper.
5th step: the bianry image after denoising is found datum line by hough conversion, and calculates according to datum line
Angle, finally carries out slant correction to bianry image, the bianry image after being corrected.
6th step: the bianry image after slant correction is carried out connected component labeling, finds all of connected region, and according to
Connected domain is screened by the size information of connected domain, is got rid of by distracter the least for area.
7th step: the connected region filtered out is carried out by positional information the coarse positioning of paper volume head.
8th step: the volume head image section of coarse positioning is carried out level, vertical direction projection, checks whether to meet volume head table
The feature of lattice.The part meeting table features is defined as rolling up the exact position of head.
9th step: according to volume head accurate position, the bianry image after slant correction is carried out cutting, extract the head figure that makes the test
Picture.
Tenth step: volume head image is carried out level, vertical direction projection, calculates the coordinate position of form.
11st step: have spherical aberration owing to high photographing instrument gathers image, in order to avoid the impact caused that distorts as far as possible, first
First image is carried out vertical segmentation, it is thus achieved that be less than the score graph table images of little topic.
12nd step: the score graph table images of every little topic is carried out the segmentation of horizontal direction, it is thus achieved that this little topic mark every
The image of one-bit digital.
13rd step: segmentation result is verified by the paper structural information according to obtaining from central server, it is ensured that figure
As segmentation is accurate.
14th step: the digital picture of each handwritten form is carried out denoising by Gassian low-pass filter.
15th step: the digital picture after denoising is carried out image enhaucament by morphologic expanding method, it is thus achieved that strengthen
After digital picture.
16th step: by being inwardly indented, the method for morphological reconstruction, is removed frame to enhanced digital picture
Process.
17th step: the digital picture removing frame is carried out numeral identification, the every problem of final acquisition by svm grader
Purpose mark.
Image processing algorithm described in the present embodiment is subdivided into four parts:
One, Image semantic classification
First the coloured image collected is processed as gray level image by gray processing, and by local auto-adaptive binaryzation or
Carry out OSTU binaryzation after edge extracting and gray level image is transferred to the bianry image of only black and white.Then to and know that image is carried out
Remove noise for the first time.Finally application hough converter technique carries out slant correction to image.
Two, volume head image zooming-out
Volume head is foregoing region to be identified.Application connected domain method carries out coarse positioning to volume head, and is had according to form
Volume head is accurately positioned by some features, is then cut out the region needing to identify.
Three, volume head form segmentation
Application level and upright projection, be partitioned into the score chart picture of the little topic of per pass.Application distortion correction and image thinning method gram
Take the spherical aberration caused when high photographing instrument gathers image.And coordinate the json string obtained from central server, i.e. paper structure
Segmentation result is verified by information, it is ensured that segmentation is correct.
Four, mark identification
With handwriting digital sample training grader before identifying, the present embodiment uses support vector machine method.To score chart
Go dry as carrying out second time, and carry out image enhaucament by operations such as Morphological scale-space and elimination frame process.Finally identify
The student number of student and the mark of per pass exercise question.System efficiency is stepped on, if topic answer of wherein filling a vacancy correctly can not be filled out for improving results
Or fill out " √ ", the full marks of topic of filling a vacancy must be divided into；If answer mistake fills out "×", zero must be divided into.
User is by " space " key on percussion keyboard or clicking on " confirming to preserve " button uploads to center service by achievement
In device.Can also check that by clicking on " checking " button the institute currently having stepped on system is fruitful, refering to Fig. 4 simultaneously.
In the present embodiment, user uses browser to sign in Result statisticses and analysis end system.System obtains from central server
Fetch data, by counting statistics, analysis result showed user with forms such as form, rectangular histogram, curve chart and word descriptions,
(the chart storehouse of a kind of pure written in JavaScript) that iconic content therein is realized by Highcharts.In main
The analytical data such as appearance includes: the statistics of average mark, item difficulty prediction and measured curve, examination diagnosis report.
Wherein the statistics of average mark is in units of class, by suing for peace every student's score in this class, then removes
With total number of persons, obtaining the average mark of this class, formula is:.T is this class of number of student,For i-th student's
Achievement.Refering to Fig. 5, different bar diagrams represents difference minimum point of class, average mark, best result, the cake chart of top-left position
Represent the distribution situation of different mark section number of student.
Whether the prediction of item difficulty and the difficulty of the gone out examination question of measured curve figure mainly scientificity meet expection.This
Carrying out scalar difficulty value by the situation of Awareness percent in embodiment, the formula calculating item difficulty value is:, wherein P
For Awareness percent., R be this topic do to number, T is answer total number of persons.
It is as follows that item difficulty prediction curve figure generates process: per pass exercise question can be carried out difficulty when that teacher going out examination question pre-
Survey, such as class's total number of persons is 50 people, teacher predict the 1st topic have 40 classmates do right, then first topic difficulty value foundation
Above-mentioned formula is just for 1-40/50=0.2.According to the method, teacher carries out difficulty value prediction to per pass exercise question in whole paper.So
After with topic number as abscissa, item difficulty value is vertical coordinate, marks corresponding coordinate points, and successively use according to topic number in coordinate
Smoothed curve is sequentially connected with coordinate points, thus obtains item difficulty prediction curve figure.
It is similar with item difficulty prediction curve figure that item difficulty measured curve figure generates process, after Data Input completes, and meeting
There is the true Awareness percent of per pass exercise question, calculate the true of per pass examination question according to real data by above-mentioned item difficulty value formula
Real difficulty value, then according to item difficulty prediction curve figure generating mode obtains item difficulty measured curve.Refering to Fig. 6.
Wherein diagnosis report includes the diagnosis to paper, rate of losing points diagnosis, the diagnosis of every student knowledge Grasping level,
Raw ability point diagnosis.
A, diagnosis to paper
Mainly include paper discrimination (whether can distinguish the student of each horizontal level), Degree of difficulty of test paper, the collocation of knowledge point
With stressing of ability point, the diagnosis to paper refers to Fig. 7.
Discrimination refers to the examination question differentiation degree to examinee's practical level.The computing formula of examination question discrimination D
(PH, PL are respectively the packet of examination question height and the Awareness percent of low packet examinee to D=PH-PL, and Awareness percent isWherein R
Do to number, T is total number of persons)
Such as in one time biological test, in 100 students, height packet is respectively arranged with 27 people, and the first topic is answered questions in the highest packet to be had
20 people, what the first topic was answered questions in low packet has 5 points, and the discrimination of this problem is:
D=PH-PL=20/27-5/27=0.55
Degree of difficulty of test paper is mainly weighed by examination question average degree of difficulty.Average degree of difficulty is the reality of all item difficulty values
Survey meansigma methods.Having the above-mentioned item difficulty value that obtains is V, and the formula of average degree of difficulty A is:；WhereinIt is i-th
Problem purpose difficulty value, q is the sum of examination question in paper.
The distribution proportion of each knowledge point is mainly listed in knowledge point collocation, checks that the distribution of knowledge point is more also
It is to give priority to.Distribution proportion mainly obtains divided by paper total score according to the mark shared by each knowledge, and distribution proportion is public
Formula is:；WhereinFor the mark shared by kth knowledge point；N is the number of knowledge point,It is i-th
Mark shared by individual knowledge point.
The distribution proportion of each ability point is mainly listed in ability point collocation, checks that the distribution of ability point is more also
It is to give priority to.Ability point distribution proportion computing formula is similar with knowledge point ratio distribution formula, i.e. with shared by each ability point
Mark divided by paper gross score.Distribution proportion formula is:；WhereinShared by kth ability point
Mark；M is the number of ability point,Mark shared by i-th ability point.
B, rate of losing points diagnose
Count whole class to lose points the higher exercise question of rate, it is proposed that teacher's emphasis explains the content of this type of exercise question.Wherein lose points
The rate problem definition more than 40% is the higher examination question of rate of losing points, and the rate formula of losing points is, wherein L is that this examination question does not has
Obtaining the number of full marks, T is this class of answer number.Refering to Fig. 8.
C, the diagnosis of student knowledge Grasping level
Per pass exercise question has corresponding knowledge point, is it is recognized that student is to knowledge according to student's per pass exercise question scoring event
The Grasping level of point.The diagnosis of student knowledge Grasping level mainly counts the scoring rate of each knowledge point.Wherein knowledge point
Scoring rate formula is:, whereinThe mark obtained by this knowledge point of i-th classmate of this class, KF is
The gross score that this knowledge point is shared in paper；T is this class of answer number.Refering to Fig. 9.
D, student ability point diagnoses
Per pass exercise question has corresponding ability point, is it is recognized that student is to ability according to student's per pass exercise question scoring event
The Grasping level of point.The diagnosis of each ability Grasping level of student mainly counts the scoring rate of each ability point.Calculate energy
Force scoring rate formula is similar with calculation knowledge point, and formula is:。Instinct for i-th classmate of this class
The mark that force is obtained, AF is the gross score that this ability point is shared in paper；T is this class of answer number.Refering to Figure 10.
The above embodiment is only the preferred embodiments of the present invention, and and non-invention possible embodiments exhaustive.
For persons skilled in the art, done any aobvious to it on the premise of without departing substantially from the principle of the invention and spirit
And the change being clear to, within all should being contemplated as falling with the claims of the present invention.
Claims (10)
1. a Students'Transcript Input, it is characterised in that include collecting unit, Data Input client, fire wall, router, in
Central server and user terminal；
The input of described collecting unit obtains paper image；The output of described collecting unit terminates described Data Input client
Respective input；Described Data Input client is connected with the Internet；
Described central server is connected with the Internet through router, fire wall；
Described user terminal is connected with the Internet.
A kind of Students'Transcript Input based on image procossing the most according to claim 1, it is characterised in that: described collection is single
Unit is scanner, high photographing instrument, photographing unit, video camera or mobile phone camera.
A kind of Students'Transcript Input based on image procossing the most according to claim 1, it is characterised in that: described achievement is recorded
Entering client is computer, panel computer or mobile phone.
A kind of Students'Transcript Input based on image procossing the most according to claim 1, it is characterised in that: genuinely convinced in described
Business device is computer；Described user terminal is computer.
5. utilize the method that the Students'Transcript Input described in claim 1 carries out image procossing, it is characterised in that include walking as follows
Rapid:
(1) training of Handwritten Digital Recognition grader is carried out: by collection and the screening of handwriting digital sample, it is thus achieved that effectively
Handwriting digital sample, the method utilizing support vector machine, train and generate identify handwriting digital svm grader；
(2) utilize described collecting unit collection with the paper original color image of scoring papery of volume head；
(3) Data Input client obtains the paper original color image of scoring papery with volume head；Data Input client
By gray processing processing method, the described paper original color image of scoring papery with volume head is converted into gray level image；
(4) described gray level image is converted into bianry image by Data Input client；
(5) Data Input client utilizes Gassian low-pass filter method that described bianry image carries out denoising for the first time, obtains once
Denoising bianry image；
(6) a described denoising bianry image is found datum line by hough conversion by Data Input client, and according to benchmark
Line computation angle of inclination, finally carries out slant correction to a denoising bianry image, obtains correcting a rear denoising bianry image；
(7) Data Input client carries out connected component labeling to correcting a rear denoising bianry image, finds all of connected region
Territory, and according to the information of connected region size, connected region is screened, distracter the least for area is got rid of, obtains
Connected region after screening；
(8) Data Input client utilizes positional information that connected region after described screening carries out the coarse positioning of paper volume head,
Head image is rolled up to coarse positioning examination papers；
(9) Data Input client carries out horizontal direction, vertical direction projection respectively to coarse positioning examination papers volume head image, checks
Whether meet volume head table features；The part meeting volume head table features is defined as rolling up the exact position of head；
(10) Data Input client carries out cutting according to the exact position of volume head to correcting a rear denoising bianry image, extracts
Go out accurately volume head image；
(11) Data Input client carries out horizontal direction, vertical direction projection respectively to accurately volume head image, calculates form
Coordinate position；
(12) spherical aberration is generally had due to collecting unit acquired image, in order to avoid the shadow caused that distorts as far as possible
Ringing, Data Input client carries out vertical segmentation to image, it is thus achieved that little topic score graph table images；
(13) Data Input client carries out the segmentation of horizontal direction to little topic score graph table images, it is thus achieved that little topic mark each
The image of bit digital；
(14) according to the paper structural information obtained from central server, segmentation result is verified by Data Input client,
Authentication image segmentation is the most accurate；
(15) Data Input client utilizes Gaussian low pass wave method that each handwriting digital image is carried out denoising, is gone
Handwritten form digital picture after making an uproar；
(16) Data Input client carries out image enhaucament by morphologic expanding method to handwritten form digital picture after making an uproar, and obtains
Obtain enhancement mode handwriting digital image；
(17) by being inwardly indented the method with morphological reconstruction, enhancement mode handwriting digital image is entered by Data Input client
Row is removed frame and is processed；
(18) the enhancement mode handwriting digital image removing frame is carried out numeral knowledge by svm grader by Data Input client
, do not finally give the fractional data information of per pass exercise question, and described fractional data information be transferred to described central server；
(19) described central server carries out data process to described fractional data information, obtains analysis processing result；
(20) described user terminal and is divided from the fractional data information of described central server acquisition per pass exercise question by the Internet
Analysis result, and fractional data information and the analysis processing result of described per pass exercise question is shown by user terminal displays device.
The method of image procossing the most according to claim 5, it is characterised in that: utilize Highcharts by described analysis
The form that reason result describes with form, rectangular histogram, curve chart or word shows user.
The method of image procossing the most according to claim 5, it is characterised in that: the data of described fractional data information process
Including average mark statistics, item difficulty prediction curve and item difficulty measured curve and examination diagnosis report.
The method of image procossing the most according to claim 7, it is characterised in that: described average mark statistics is with class as list
Position, by suing for peace every student's score in described class, then divided by total number of persons, obtains the average mark of this class, expresses
Formula is:
；Wherein, T is this class of answer number,Achievement for i-th student.
The method of image procossing the most according to claim 7, it is characterised in that: described item difficulty prediction curve and examination question
It is as follows that difficulty measured curve generates method:
The formula calculating item difficulty value is:, wherein P is examination question Awareness percent,, R be this examination question do to
Number, T is for participating in answer class size；
With topic number as abscissa, item difficulty value is vertical coordinate, marks corresponding coordinate points in coordinate, and according to topic priority
It is sequentially connected with coordinate points with smoothed curve, i.e. obtains item difficulty prediction curve；
After Data Input completes, the true Awareness percent of per pass examination question is RP, calculates per pass examination question true difficulty value RV, expression formula
For；With topic number as abscissa, the true difficulty value of examination question is vertical coordinate, marks corresponding coordinate in coordinate
Point, and it is sequentially connected with coordinate points according to topic priority smoothed curve, i.e. obtain item difficulty measured curve.
The method of image procossing the most according to claim 7, it is characterised in that: described examination diagnosis report includes examination
The diagnosis of volume, rate of losing points diagnosis, the diagnosis of every student knowledge Grasping level, the diagnosis of student ability point；
A, diagnosis to paper include for distinguishing the examination question discrimination of horizontal student at all levels, Degree of difficulty of test paper, knowledge point
Collocation stresses with ability point；
Examination question discrimination refers to the examination question differentiation degree to examinee's practical level；The computing formula of examination question discrimination is
D=PH-PL
Wherein D is examination question discrimination, and PH, PL are respectively the packet of examination question height and the Awareness percent of low packet examinee, and high packet is just answered
Rate formula is:, wherein RH height packet do to number, TH is high packet number；Low packet Awareness percent formula is:, wherein the low packet of RL do to number, TL is low packet number；
Described Degree of difficulty of test paper is weighed by examination question average degree of difficulty；Described average degree of difficulty is all item difficulty values
Actual measurement meansigma methods；
The formula of average degree of difficulty A is:；WhereinBeing the i-th problem purpose difficulty value, q is examination question in paper
Sum；
Knowledge point collocation is the distribution proportion listing each knowledge point, checks that the distribution of knowledge point is the most averagely or side
Weight；Distribution proportion is to obtain divided by paper total score according to the mark shared by each knowledge, and distribution proportion formula is:；WhereinFor the mark shared by kth knowledge point；N is the number of knowledge point,For i-th knowledge
The shared mark of point；
Ability point collocation is the distribution proportion listing each ability point, checks that the distribution of ability point is the most averagely or side
Weight；Ability point distribution proportion computing formula is similar with knowledge point ratio distribution formula, i.e. removes with the mark shared by each ability point
With paper gross score；Distribution proportion formula is:；WhereinFor the mark shared by kth ability point；m
For the number of ability point,Mark shared by i-th ability point；
The purpose of b, rate of losing points diagnosis is to count whole class to lose points the higher exercise question of rate, and advises that teacher's emphasis explains this
The content of class exercise question；The rate of wherein the losing points problem definition more than 40% is the higher examination question of rate of losing points, and the rate formula of losing points is, wherein L is the number that this examination question there is no full marks, and T is this class of answer number；
C, the diagnosis of student knowledge Grasping level
Per pass exercise question has corresponding knowledge point, is it is recognized that student is to knowledge according to student's per pass exercise question scoring event
The Grasping level of point；The diagnosis of student knowledge Grasping level is the scoring rate counting each knowledge point；Wherein knowledge point score
Rate formula is:, whereinThe mark obtained by this knowledge point of i-th classmate of this class, KF knows for this
Know the gross score that point is shared in paper；T is this class of answer number；
D, student ability point diagnoses
Per pass exercise question has corresponding ability point, is it is recognized that student is to ability according to student's per pass exercise question scoring event
The Grasping level of point；The diagnosis of each ability Grasping level of student is the scoring rate counting each ability point；Computing capability point
Scoring rate formula is:；
The mark obtained by this ability point of i-th classmate of this class, AF is the total score that this ability point is shared in paper
Number；T is this class of answer number.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610186428.9A CN105869085A (en) | 2016-03-29 | 2016-03-29 | Transcript inputting system and method for processing images |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610186428.9A CN105869085A (en) | 2016-03-29 | 2016-03-29 | Transcript inputting system and method for processing images |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105869085A true CN105869085A (en) | 2016-08-17 |
Family
ID=56625165
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610186428.9A Pending CN105869085A (en) | 2016-03-29 | 2016-03-29 | Transcript inputting system and method for processing images |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105869085A (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106503747A (en) * | 2016-10-28 | 2017-03-15 | 西安夫子电子科技研究院有限公司 | A kind of image recognition statistical analysis system |
| CN107239904A (en) * | 2017-06-06 | 2017-10-10 | 合肥创旗信息科技有限公司 | A kind of online PE testing platform |
| CN107358150A (en) * | 2017-06-01 | 2017-11-17 | 深圳赛飞百步印社科技有限公司 | Object frame identification method, device and high photographing instrument |
| CN107506762A (en) * | 2017-09-01 | 2017-12-22 | 昆山中骏博研互联网科技有限公司 | A kind of achievement method for automatically inputting based on graphical analysis |
| CN107680046A (en) * | 2017-07-25 | 2018-02-09 | 平安科技（深圳）有限公司 | Image rectification method, device, storage medium and computer equipment |
| CN107862230A (en) * | 2017-11-14 | 2018-03-30 | 安徽虚境科技有限公司 | A kind of grade determines and input method |
| CN107908752A (en) * | 2017-11-18 | 2018-04-13 | 曲阜师范大学 | A kind of paper achievement intelligent acquisition and analysis system and method |
| CN108564505A (en) * | 2018-04-23 | 2018-09-21 | 温州市鹿城区中津先进科技研究院 | One kind is based on classroom performance data and achievement data correlation analysis system |
| CN108614833A (en) * | 2016-12-12 | 2018-10-02 | 英业达科技有限公司 | Analysis suitable for paper answer and statistical system and its method |
| CN109034737A (en) * | 2018-07-12 | 2018-12-18 | 方碧水 | A kind of student's PIM method based on Internet of Things |
| CN110413789A (en) * | 2019-07-31 | 2019-11-05 | 广西师范大学 | A kind of exercise automatic classification method based on SVM |
| CN111083434A (en) * | 2019-05-15 | 2020-04-28 | 广东小天才科技有限公司 | Method for starting dictation detection and electronic equipment |
| CN114489406A (en) * | 2022-01-30 | 2022-05-13 | 广州宏途数字科技有限公司 | Information input method and system for teaching service |
| CN114936809A (en) * | 2022-07-22 | 2022-08-23 | 山东悦知教育科技有限公司 | Examination score data processing system and terminal based on big data |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101038626A (en) * | 2007-04-25 | 2007-09-19 | 上海大学 | Method and device for recognizing test paper score |
| CN101315668A (en) * | 2008-07-01 | 2008-12-03 | 上海大学 | Automatic detection method for test paper form |
| CN101685482A (en) * | 2009-08-04 | 2010-03-31 | 上海心意答电子科技有限公司 | Electric marking system capable of automatically processing marking results and method thereof |
| CN203217591U (en) * | 2013-04-02 | 2013-09-25 | 大连交通大学 | Intelligent test-paper score statistical system |
| CN104732184A (en) * | 2015-03-24 | 2015-06-24 | 朱永峰 | Art department student exam score inputting system |
| CN104751187A (en) * | 2015-04-14 | 2015-07-01 | 山西科达自控股份有限公司 | Automatic meter-reading image recognition method |
| CN104850663A (en) * | 2015-06-09 | 2015-08-19 | 苏州点通教育科技有限公司 | System and method for inputting and analyzing test scores |
-
2016
- 2016-03-29 CN CN201610186428.9A patent/CN105869085A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101038626A (en) * | 2007-04-25 | 2007-09-19 | 上海大学 | Method and device for recognizing test paper score |
| CN101315668A (en) * | 2008-07-01 | 2008-12-03 | 上海大学 | Automatic detection method for test paper form |
| CN101685482A (en) * | 2009-08-04 | 2010-03-31 | 上海心意答电子科技有限公司 | Electric marking system capable of automatically processing marking results and method thereof |
| CN203217591U (en) * | 2013-04-02 | 2013-09-25 | 大连交通大学 | Intelligent test-paper score statistical system |
| CN104732184A (en) * | 2015-03-24 | 2015-06-24 | 朱永峰 | Art department student exam score inputting system |
| CN104751187A (en) * | 2015-04-14 | 2015-07-01 | 山西科达自控股份有限公司 | Automatic meter-reading image recognition method |
| CN104850663A (en) * | 2015-06-09 | 2015-08-19 | 苏州点通教育科技有限公司 | System and method for inputting and analyzing test scores |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106503747A (en) * | 2016-10-28 | 2017-03-15 | 西安夫子电子科技研究院有限公司 | A kind of image recognition statistical analysis system |
| CN108614833A (en) * | 2016-12-12 | 2018-10-02 | 英业达科技有限公司 | Analysis suitable for paper answer and statistical system and its method |
| CN107358150A (en) * | 2017-06-01 | 2017-11-17 | 深圳赛飞百步印社科技有限公司 | Object frame identification method, device and high photographing instrument |
| CN107239904A (en) * | 2017-06-06 | 2017-10-10 | 合肥创旗信息科技有限公司 | A kind of online PE testing platform |
| CN107680046A (en) * | 2017-07-25 | 2018-02-09 | 平安科技（深圳）有限公司 | Image rectification method, device, storage medium and computer equipment |
| WO2019019383A1 (en) * | 2017-07-25 | 2019-01-31 | 平安科技（深圳）有限公司 | Image correction method and apparatus, storage medium and computer device |
| CN107506762A (en) * | 2017-09-01 | 2017-12-22 | 昆山中骏博研互联网科技有限公司 | A kind of achievement method for automatically inputting based on graphical analysis |
| CN107506762B (en) * | 2017-09-01 | 2020-03-20 | 昆山中骏博研互联网科技有限公司 | Score automatic input method based on image analysis |
| CN107862230A (en) * | 2017-11-14 | 2018-03-30 | 安徽虚境科技有限公司 | A kind of grade determines and input method |
| CN107908752A (en) * | 2017-11-18 | 2018-04-13 | 曲阜师范大学 | A kind of paper achievement intelligent acquisition and analysis system and method |
| CN108564505A (en) * | 2018-04-23 | 2018-09-21 | 温州市鹿城区中津先进科技研究院 | One kind is based on classroom performance data and achievement data correlation analysis system |
| CN109034737A (en) * | 2018-07-12 | 2018-12-18 | 方碧水 | A kind of student's PIM method based on Internet of Things |
| CN111083434A (en) * | 2019-05-15 | 2020-04-28 | 广东小天才科技有限公司 | Method for starting dictation detection and electronic equipment |
| CN110413789A (en) * | 2019-07-31 | 2019-11-05 | 广西师范大学 | A kind of exercise automatic classification method based on SVM |
| CN114489406A (en) * | 2022-01-30 | 2022-05-13 | 广州宏途数字科技有限公司 | Information input method and system for teaching service |
| CN114489406B (en) * | 2022-01-30 | 2023-03-10 | 广州宏途数字科技有限公司 | Information input method and system for teaching service |
| CN114936809A (en) * | 2022-07-22 | 2022-08-23 | 山东悦知教育科技有限公司 | Examination score data processing system and terminal based on big data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105869085A (en) | Transcript inputting system and method for processing images | |
| JP6892558B2 (en) | Theological assistance method and the theological assistance system that adopts the method | |
| CN104881662B (en) | A kind of single image pedestrian detection method | |
| CN107180239A (en) | Line of text recognition methods and system | |
| CN104077577A (en) | Trademark detection method based on convolutional neural network | |
| CN110135282B (en) | Examinee return plagiarism cheating detection method based on deep convolutional neural network model | |
| CN110472494A (en) | Face feature extracts model training method, facial feature extraction method, device, equipment and storage medium | |
| CN105825189A (en) | Device for automatically analyzing attendance rate and class concentration degree of college students | |
| CN106446882A (en) | method for intelligently marking paper with trace left based on 8-character code | |
| CN107480607B (en) | Method for detecting and positioning standing face in intelligent recording and broadcasting system | |
| CN112183238B (en) | Remote education attention detection method and system | |
| US10037708B2 (en) | Method and system for analyzing exam-taking behavior and improving exam-taking skills | |
| CN106709438A (en) | Method for collecting statistics of number of people based on video conference | |
| CN110175534A (en) | Teaching assisting system based on multitask concatenated convolutional neural network | |
| CN110147736A (en) | A kind of repetition anomaly detection method based on posture | |
| CN108876195A (en) | A kind of intelligentized teachers ' teaching quality evaluating system | |
| CN110543811A (en) | non-cooperation type examination person management method and system based on deep learning | |
| CN103455826B (en) | Efficient matching kernel body detection method based on rapid robustness characteristics | |
| CN111444389A (en) | Conference video analysis method and system based on target detection | |
| CN112396635A (en) | Multi-target detection method based on multiple devices in complex environment | |
| CN111611854B (en) | Classroom condition evaluation method based on pattern recognition | |
| CN109598681A (en) | The reference-free quality evaluation method of image after a kind of symmetrical Tangka repairs | |
| CN108428073A (en) | A kind of intelligent evaluation system for teachers ' teaching quality | |
| CN114677644A (en) | Student seating distribution identification method and system based on classroom monitoring video | |
| CN113989608A (en) | Student experiment classroom behavior identification method based on top vision |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20160817 |
|
| RJ01 | Rejection of invention patent application after publication |